Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING

A First Assessment of the P-SBAS DInSAR

Algorithm Performances Within a Cloud
Computing Environment
Ivana Zinno, Stefano Elefante, Lorenzo Mossucca, Claudio De Luca, Michele Manunta, Olivier Terzo,
Riccardo Lanari, Fellow, IEEE, and Francesco Casu

AbstractWe present in this work a first performance assessment of the Parallel Small BAseline Subset (P-SBAS) algorithm,
for the generation of Differential Synthetic Aperture Radar (SAR)
Interferometry (DInSAR) deformation maps and time series,
which has been migrated to a Cloud Computing (CC) environment. In particular, we investigate the scalable performances
of the P-SBAS algorithm by processing a selected ENVISAT
ASAR image time series, which we use as a benchmark, and by
exploiting the Amazon Web Services (AWS) CC platform. The
presented analysis shows a very good match between the theoretical and experimental P-SBAS performances achieved within
the CC environment. Moreover, the obtained results demonstrate
that the implemented P-SBAS Cloud migration is able to process
ENVISAT SAR image time series in short times (less than 7 h)
and at low costs (about USD 200). The P-SBAS Cloud scalable
performances are also compared to those achieved by exploiting
an in-house High Performance Computing (HPC) cluster, showing that nearly no overhead is introduced by the presented Cloud
solution. As a further outcome, the performed analysis allows us
to identify the major bottlenecks that can hamper the P-SBAS
performances within a CC environment, in the perspective of processing very huge SAR data flows such as those coming from the
existing COSMO-SkyMed or the upcoming SENTINEL-1 constellation. This work represents a relevant step toward the challenging
Earth Observation scenario focused on the joint exploitation of
advanced DInSAR techniques and CC environments for the massive processing of Big SAR Data.
Index TermsBig data, Cloud Computing (CC), Differential
Synthetic Aperture Radar (SAR) Interferometry (DInSAR), Earth
surface deformation, Parallel Small BAseline Subset (P-SBAS).
Manuscript received November 06, 2014; revised February 13, 2015;
accepted April 13, 2015. This work was supported in part by the Italian
Ministry of University and Research (MIUR) under the project Progetto
Bandiera RITMARE, and in part by the Italian Civil Defence Department
(DPC) of the Prime Ministers Office. This work has been carried out
through the I-AMICA (Infrastructure of High Technology for Environmental
and Climate MonitoringPONa3_00363) project of Structural improvement
financed under the National Operational Programme (NOP) for Research and
Competitiveness 20072013, cofunded with European Regional Development
Fund (ERDF) and National resources. The ENVISAT SAR data have been provided by the European Space Agency through the Virtual Archive 4. The DEM
of the investigated zone was acquired through the SRTM archive.
I. Zinno, S. Elefante, C. De Luca, M. Manunta, R. Lanari, and F. Casu
are with the Istituto per il Rilevamento Elettromagnetico dellAmbiente
(IREA), Consiglio Nazionale delle Ricerche, Napoli 80124, Italy (e-mail:
casu.f@irea.cnr.it).
C. De Luca is also with the Department of Electrical Engineering and
Information Technology, University of Naples Federico II, Napoli, Italy.
L. Mossucca and O. Terzo are with the Advanced Computing and
Electromagnetics (ACE), Istituto Superiore Mario Boella, Torino 10138, Italy.
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/JSTARS.2015.2426054

I. I NTRODUCTION

DVANCED differential synthetic aperture radar (SAR)

interferometry (DInSAR) usually identifies a set of algorithms, tools, and methodologies for the generation of Earths
surface deformation maps and time series computed from a
sequence of multitemporal differential SAR interferograms [1].
A widely used advanced DInSAR approach is the technique
named Small BAseline Subset (SBAS) [2] which allows to produce line-of-sight (LOS)-projected mean deformation velocity
maps and corresponding displacement time series by exploiting
interferograms characterized by a small temporal and/or spatial separation (baseline) between the acquisition orbits. The
SBAS algorithm has proven its effectiveness to detect ground
displacements with millimeters accuracy [3] in different scenarios, such as volcanoes, tectonics, landslides, anthropogenicinduced land motions [4][7] and it is capable to perform
analyses at different spatial scales [8] and with multisensor
data [9], [10].
The SBAS algorithm, and more generally the advanced
DInSAR techniques, found their success on the large availability of SAR data archives acquired over time by several
satellite systems. Indeed, the current radar Earth Observation
(EO) scenario takes advantage of the widely diffused longterm C-band ESA (e.g., ERS-1, ERS-2, and ENVISAT) and
Canadian (RADARSAT-1/2) SAR data archives, which have
been acquired during the last 20 years, as well as of data
sequences provided by the X-band generation SAR sensors,
such as the COSMO-SkyMed (CSK) and TerraSAR-X (TSX)
constellations. Moreover, a massive and ever increasing data
flow will be further supplied by the recently launched (April
2014) Copernicus (European Union) SENTINEL-1A SAR
satellite, which will also be paired during 2016 with the
SENTINEL-1B twin system that will allow halving the constellation revisit time (from 12 to 6 days) [11]. With the
SENTINEL-1 era, new SAR data relevant to extended areas on
Earth will be soon publically available, thanks to the free and
open access data policy adopted by the Copernicus program.
Moreover, the SENTINEL-1 data will be collected on land by
using the TOPS SAR mode that has been specifically designed
for DInSAR and advanced DInSAR applications [12], [13].
In this context, the massive exploitation of these Big SAR
data archives for the generation of advanced DInSAR products will open new research perspectives to understand Earths
surface deformation dynamics at global scale.

1939-1404 2015 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution
requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
2

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING

However, the accomplishment of this task requires not only

large and very high computing resources to process the existing and upcoming huge SAR data amounts within short time
frames, but also efficient algorithms able to effectively exploit
the available computing facilities.
To provide a contribution toward this direction, a parallel version of the SBAS algorithm, namely parallel SBAS (P-SBAS),
has been recently proposed [14]. P-SBAS permits to generate,
in an automatic and unsupervised manner, advanced DInSAR
products by taking full benefit from parallel computing architectures, such as cluster and GRID infrastructures. P-SBAS has
been extensively tested by exploiting in-house processing facilities achieving promising results in terms of both scalability and
efficiency [14][16]. However, it is worth noting that, even if inhouse solutions can provide high computing performances, they
can represent a bottleneck due to their intrinsic limited resource
availability. Therefore, they cannot be suited to properly face
the massive processing that will be inevitably required by the
expected huge SAR data flow, particularly when global scale
analyses are concerned. Moreover, in-house High Performance
Computing (HPC) infrastructures can be very expensive in
terms of procurement, maintenance, and upgrading.
The use of Cloud Computing (CC) environments represents
a promising solution to overcome the above-mentioned limitations and this is the rationale why they are becoming more and
more diffused in EO scenarios [17][19]. Indeed, CC provides
highly scalable and flexible architectures that are, in general,
computationally very efficient and less expensive with respect
to in-house solutions. In addition, CC can be extremely helpful
for both resources optimization and performance improvements
due to its possibility to build up customized computing infrastructures. Moreover, the increasing availability of public Cloud
environments [20][22], and their relative simplicity to use,
thanks to advanced application programming interfaces (API)
and web-based tools, is further pushing toward the use of such
a technology also in scientific applications [19], [23], [24]. In
this context, the migration of scientific applications to CC environments is, therefore, a key issue, with particular reference
to advanced DInSAR algorithms, because a significant initial
effort and a deep preliminary analysis of the specific algorithm
which has to be cloudified could be required.
We present in this work a first performance assessment of
the P-SBAS algorithm that has been migrated to a public CC
environment. In particular, the goal of our study is to evaluate
the P-SBAS scalable performances achieved within a CC environment as well as to identify the major inefficiency sources
that can hamper such performances. To this aim, we select the
proper Cloud migration approach that allows the full exploitation of the P-SBAS parallelization strategy. Moreover, due to
the high P-SBAS computational requirements, we evaluate the
most appropriate Cloud resources configuration, in terms of
both instances and storage. An extensive analysis is carried
out by processing a selected SAR images time series acquired
(from ascending orbits) by the ENVISAT ASAR sensor over the
Napoli bay area and by exploiting the Amazon Web Services
(AWS) Cloud platform. The obtained P-SBAS scalable performances are also compared to those achieved by exploiting an
in-house, dedicated, HPC cluster.

The paper is organized as follows. In Section II, a concise description of the P-SBAS processing chain is provided,
which aims at recalling the main processing steps. Section III
describes how the entire P-SBAS processing chain has been
migrated to the selected CC environment. Section IV is dedicated to the experimental framework that includes the scalable
performance study, which has been performed by exploiting both a dedicated cluster and the AWS Cloud, as well
as the analysis of the processing times and costs. Finally, in
Section V, some concluding remarks and further developments
are thoroughly discussed.
II. P-SBAS D ESCRIPTION
This section focuses on providing a concise but informative description of the P-SBAS processing chain which has
already been thoroughly discussed in [14]. It aims at recalling the P-SBAS major processing steps in terms of main tasks,
implemented procedures and computational challenges that will
be addressed by showing CPU usage, RAM occupation, and
Input/Output (I/O) transfer requirements.
The P-SBAS solution has been designed by carefully taking into account several different conceptual aspects, such
as data dependencies, task partitioning, inherent granularity,
scheduling policy, load unbalancing, in order to optimize the
usage of CPU, RAM, and I/O resources. Moreover, the heterogeneous nature of the computational algorithms that are
comprised within the SBAS processing chain has strived the
P-SBAS design toward the employment of proper parallelization strategies that depend on the algorithmic structure of the
considered specific processing step [14], [18]. Furthermore,
P-SBAS has been designed in a manner that allows us to
take advantage of both multinode and multicore architectures
and therefore two-parallelization levels have been employed:
process and thread. The former considers a coarse/medium
granularity-based approach and has been mainly applied to the
whole processing chain, while the latter relies on a fine-grained
parallelization. This strategy has been implemented both for the
most computing-intensive operations, to optimize CPU usage
through multithreading programming, as well as for highly
RAM demanding algorithms, to reduce memory occupation by
applying a data parallelism strategy [14].
The block diagram of the P-SBAS processing chain is shown
in Fig. 1; in this scheme, the steps depicted by blue blocks
represent the jobs that are parallel executed by simultaneously running on different nodes, while black blocks represent
steps that are intrinsically sequential. Moreover, dashed line
blocks describe the steps that are multithreaded programmed. In
Table I, instead, the main characteristics of each step of Fig. 1,
in terms of CPU and RAM usage as well as I/O operations, are
briefly summarized in Table I.
In the following, a conceptual description of the P-SBAS
processing chain will be briefly addressed. It is worth noting
that such a processing chain has been designed to analyze the
majority of SAR data available through the different spaceborne
systems (ERS-1/2, ENVISAT, COSMO-SkyMed, TerraSAR-X,
ALOS-1/2, and RADARSAT-1/2). Moreover, it is also robust
with respect to possible system failures (e.g., ERS-2 gyroscope

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
ZINNO et al.: FIRST ASSESSMENT OF P-SBAS DINSAR ALGORITHM PERFORMANCES

Fig. 1. P-SBAS workflow. Black and blue blocks represent sequential and parallel (from a process-level perspective) processing steps, respectively. Dashed line
blocks represent multithreading programmed processing steps.

failure event) and ancillary data inaccuracies (e.g., inaccuracy

of orbital information). Step A implements the SAR data focusing operation and consists in transforming the radar raw data
into microwave images, often referred to as single look complex (SLC) images. This step has not only a high computational
burden because it performs two-dimensional FFTs on large
matrices [25] but has also a remarkably significant data flow as
it involves many I/O operations (see Table I). Step B performs
the digital elevation model (DEM) conversion that consists in
referring the elevation profile of the area under consideration
with respect to the SAR coordinate reference system [26]. In
this step, data matrices that are typically of the order of several
GBytes are processed and therefore this step requires a significant high amount of memory and I/O operations, as highlighted
in Table I.
Within step C, the SAR image coregistration operation is carried out to refer all the SLCs to the radar geometry of a selected
reference master, via an interpolation procedure [27]. The
main limitation of this step is represented by the intensive I/O
access and large amount of available memory that are required.
Subsequently, step D performs the identification of interferometric data pairs that are required for the subsequent
coregistration refinement performed within the following step
E, in which first possible residual subpixel rigid shifts are evaluated, which are, afterward, inverted and used to resample the
whole image stack. This step is computationally high demanding because of the applied resampling method that is based on
FFTs, which are calculated on large complex matrices.
Within step F, some parameters related to the interferometric
data pairs, which needs to be included in the following of the
processing chain are selected [25], [26], [28]. Having already
both the coregistered images (step C) and the DEM referred to

a common radar geometry (as output of step B), the differential

interferograms and the corresponding spatial coherence map
are generated through step G [28]. These operations require
a large amount of RAM memory usage as they are carried
out at the SAR images full spatial resolution and in the complex domain. However, the final interferograms are stored in
low resolution mode through a complex spatial average (multilook) operation [25]. This procedure, undertaken to mitigate
the decorrelation noise affecting the DInSAR interferograms,
also drastically reduces the sizes of the final outputs, while the
intermediate products remains at full resolution.
The modulo-2 restricted phase of each computed multilook
interferogram needs afterward to be unwrapped to retrieve the
original phase [25]. This procedure is carried out in step H and
I by applying the extended minimum cost flow (EMCF) phase
unwrapping (PhU) algorithm [29].
The phase unwrapping step is one of the most demanding in
terms of memory and CPU usage. Indeed, it deals with wrapped
and unwrapped interferogram stacks that are represented by
three-dimensional matrices.
A pixel-based inversion of the unwrapped phase system of
equations is, afterward, carried out (step J) to retrieve the final
deformation time-series. Moreover, in step K, the estimation
of possible residual phase artifacts often referred to as orbital
phase ramps is undertaken by exploiting interferograms. Such
phase ramps, that are due to possible orbital inaccuracies,
are removed from the wrapped interferograms, implying that
another PhU step on the orbital error free interferograms has
to be performed (second run of steps H, I, and J of Fig. 1).
After executing temporal coherence estimation [29] used for the
coherent pixel selection, block L provides the final deformation
time-series.

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
4

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING

TABLE I
P-SBAS A LGORITHM R ESOURCE R EQUIREMENTS

Reference values for a standard dataset (64 ENVISAT SAR images, see Section IV). CPU: medium (<100%),
high (100%300%), very high (>300%); RAM: low (<1 GB), medium (110 GB), high (>10 GB); I/O: medium
(<10 GB), high (tens of GB), very high (hundreds of GB).

In the following, the main critical issues concerning the

porting of the P-SBAS algorithm within CC environments are
summarized.
First, P-SBAS requires that the execution of the different
steps needs to be performed following a well-defined specific order. For instance, SAR image registration can occur
only once the focusing operation has been accomplished, while
the PhU can be executed only after the differential interferograms sequence generation has been achieved. This rationale
motivates why the generated data and products of the SBAS
processing chain are highly interconnected and, as a consequence of this data dependency, the joint processing of different
step outputs needs at some point to be performed.
Second, many steps of the chain, such as the raw data
focusing (step A), image coregistration (steps C and E), interferogram generation (step G) and the unwrapping steps (H and
I), are characterized by a large amount of RAM usage, thus
implying a proper resources allocation design. Finally, very
large amount of data are also involved during the SBAS processing and, therefore, data transfer and I/O issues are two
critical factors to be carefully taken into account. The usage of
ad hoc storage capability and network facility is envisaged to be
a viable solution for a proper P-SBAS algorithm exploitation.

III. C LOUD I MPLEMENTATION

Among the different possibilities for integrating an existing
application in a CC environment that comprise either adaptation
or redesign of the application components [30], we followed the
approach of migrating the whole P-SBAS processing chain to
the Cloud. This choice has been founded on the criterion of
applying a minimum invasiveness policy to the existing application implementation. This is a basic example of migration
to the Cloud, where the application is encapsulated in virtual
machines (VMs) which run within the CC environment [30].
Thanks to the use of VMs, several benefits can be exploited
such as: 1) security and isolation, as services can run in
Cloud environment totally independent from each other; 2) ease
administration and management, due to the common virtualization layer; 3) disaster recovery, as VMs can be launched in
few minutes and furthermore can be cloned and migrated to
different locations; and 4) high reliability and load balancing
optimization.
The Cloud architecture, depicted in Fig. 2, has been designed
taking into account the P-SBAS algorithm analysis carried out
in the previous section and consists of several worker nodes
(WNs) connected through the network to a master node (MN),
which is also used as WN.

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
ZINNO et al.: FIRST ASSESSMENT OF P-SBAS DINSAR ALGORITHM PERFORMANCES

Fig. 2. Cloud architecture for P-SBAS analysis constituted of several WNs which include a MN and a common storage volume in a RAID 0 configuration. Each
component is located in the Amazon Web Services Cloud.

As mentioned in the previous section, P-SBAS rationale provides that in many steps of the chain the joint processing of
the outputs generated by previous steps needs to be performed.
Hence, a common storage is required and therefore a network
file system (NFS) has been adopted [31].
Moreover, the majority of the P-SBAS algorithms are developed in the Exelis Interactive Data Language (IDL) [32]: a
programming language that is widely used by the scientists who
develop algorithms for SAR and DInSAR data processing [14],
[33]. IDL is a commercial software and, therefore, each VM
running the application requires a license. Hence, an interconnection layer between the IDL License Server (located at the
site of CNR-IREA institution) and VMs (located on Cloud) has
been implemented. This layer allows us to connect end-points
and satisfies the firewall policies adopted by the CNR-IREA
institution and the Cloud provider.
The used CC platform is hosted in the Amazon Elastic
Compute Cloud (EC2), an infrastructure as a service (IaaS) that
is part of AWS Cloud; EC2 has been chosen because it is currently a feature-rich, stable and commercial public Cloud [20].
Moreover, a web service through which users can configure and
instantiate a VM image is also available. AWS adopts a virtualization technology that permits to flexibly configure VM
instances allowing users to fully set up features such as the
number of CPUs cores, the processor type, the memory, the
I/O performance, etc. In addition, the operating system and the
software that runs on the VM can be customized by the user.
In particular, a Virtual Private Cloud (VPC) has been implemented that is a logically isolated section of the AWS in which
resources can be launched in a completely defined virtual network. The easy customization of the network configuration
allows users to fully control the virtual networking environment, through IP address range selection, subnet creation,

routing tables configuration, network gateways, and multiple

layers of security (security groups and network access control
lists). Thanks to this customization, the interconnection layer
for IDL License Server linking and SAR data uploading has
been configured with predefined rules to ensure a secure connection between end points and allow only authorized users to
share data. Furthermore, the VM images have been configured
by exploiting Linux operating system, as required by the PSBAS algorithm and all the software and libraries needed for
the processing (IDL, C++, etc.) have been installed.
Finally, it is worth noting that the Cloud resources provisioning and configuration phases are automatically performed
through dedicated scripts written in Linux Bash. Accordingly,
the P-SBAS deployment in a different Cloud environment
should be rather easy and quick to be carried out because it
requires only slight modifications to such scripts in order to
exploit the specific Cloud API.
IV. E XPERIMENTAL F RAMEWORK
The aim of this section is to evaluate the scalable performances of the P-SBAS processing chain within a CC environment and compare them to the corresponding results achieved
on an in-house cluster located at CNR-IREA institute premises.
Such an analysis has been performed by migrating the existing
overall P-SBAS application to the selected AWS CC platform and, besides, it has been aimed at identifying the major
inefficiency sources that can hamper its performances. It is
worth noting that a one-to-one performance comparison is not
fully achievable due to the different resources and architectures underlying the Cloud and in-house computing infrastructures. However, the specific solution implemented within the
Cloud environment has been targeted to provide the computing

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
6

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING

TABLE II
CNR-IREA C LUSTER C ONFIGURATION

resources as much comparable as possible between the two used

infrastructures, as discussed in the following.
A. Computational Platform
The experimental analysis, as above mentioned, has been carried out by exploiting both an in-house HPC cluster which is
located at CNR-IREA premises and the public AWS Cloud.
The CNR-IREA cluster consists of 16 nodes, each one is
equipped with two CPUs (eight-core 2.6 GHz Intel Xeon E52670) and 384 GB of RAM (see Table II). The cluster has
a storage shared among different nodes that is implemented
through NFS and employs a 56-Gb/s InfiniBand interconnection. In particular, each processing node is equipped with a
direct attached storage (DAS) system in a RAID 5 configuration which ensures a disk access bandwidth of approximately
300 MB/s.
Concerning the AWS Cloud architecture described in the previous section, it has been implemented by selecting, among the
available ones, the c3.8xlarge EC2 instance that has been used
for both master and working nodes (see Table III). Indeed, such
an instance is equipped with a processor similar to the one that
is present within the in-house cluster nodes (high-frequency
Intel Xeon E5-2680 v2 processor). It is also worth noting that
AWS describes the EC2 instance computing performances in
terms of virtual CPUs that are not easily referable to physical
CPUs and cores. To make as much as possible comparable the
cluster and Cloud analyses, we treated a single AWS instance as
a single node and we forced the algorithm multithreading parts
to run using the same fixed number of cores in both the cluster
and the Cloud cases. Moreover, the selected c3.8xlarge instance
has got an amount of RAM sufficient to run the P-SBAS application without incurring in page faulting and is connected with
a 10-Gb/s network bandwidth which is the largest available
within AWS. The choice of an instance with such a very high
network bandwidth has been driven to guarantee high data
transfer and I/O performances as much as possible similar to
those of the in-house cluster. As a matter of fact, one of the
major critical issues of the P-SBAS algorithm is the very large
amount of data (in terms of inputs, intermediate products and
final results) that are read/written during the overall processing
into the common NFS storage; therefore, particular attention
has been addressed to the choice of the storage volume type as
well as of the network which allows taking full advantage of
the storage I/O performance. AWS gives users the possibility to

TABLE III
AWS I NSTANCE T YPE C ONFIGURATION

select among different types of storage according to the specific

processing needs; in this case the Amazon Elastic Block Store
(EBS) has been employed which provides persistent block-level
storage volumes to use together with EC2 instances. In particular, the provisioned IOPS (SSD) EBS volume type has
been selected since it is designed to meet the needs of I/O
intensive workloads and can provision up to 4000 IOPS (I/O
operation per second) which corresponds to about 128 MB/s
of throughput [20]. Moreover, to improve this latter parameter,
two EBS volumes have been selected and configured in RAID
0 (Striping) that allows summing the throughput of the two volumes within it, thus achieving in our case a total bandwidth of
about 256 MB/s. The 10-Gb/s network is able to support both
such a read/write bandwidth and the communications among
different nodes and storage (up to a certain number of exploited
nodes as widely discussed in the following).
B. Parallel Computing Metrics and Exploited Dataset
In order to quantitatively evaluate the scalable performances
of the P-SBAS processing chain, appropriate metrics were
adopted. Let N be the number of computing nodes used to
solve a problem and T1 be the execution time of the sequential implementation to solve the same problem, the speedup SN
of a parallel program with parallel execution time TN is defined
as [34]
SN =

T1
.
TN

(1)

Accordingly, the speedup is a metric that compares the

parallel and sequential execution times.
An alternative performance measure for a parallel program is
the efficiency [34], [35]
=

SN
N

(2)

which is a measure of the speedup achieved per computing

node. It should be noted that an ideal speedup corresponds to
a unitary efficiency.
It is worth highlighting that the parallel performance of the
P-SBAS algorithm has already been discussed in [14]; it was
shown that the major source of efficiency loss is given by
the amount of sequential part within the P-SBAS processing

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
ZINNO et al.: FIRST ASSESSMENT OF P-SBAS DINSAR ALGORITHM PERFORMANCES

Fig. 3. Mean deformation velocity map of the Napoli Bay area, generated via the P-SBAS processing on AWS Cloud. The graph of the displacement time series
relevant to two pixels located in the area of maximum deformation of Campi Flegrei and in the Vomero residential hill are also shown.

chain which, even if liable to a remarkable further reduction,

remains essentially noneliminable due to the complex nature of
the algorithm.
To quantitatively assess the effect of the serial parts of the
algorithm on the attainable speedup, the well-known Amdahls
law is hereafter introduced [35]
A
=
SN

1
,
1 fs
fs +
N

0 fs 1

(3)

where fs is the parallel program fraction that has been sequentially executed (sequential fraction) [35]. It is also worth mentioning that the formulation (3) of Amdahls law does not take
into account either the load unbalancing or the data transfer
overhead.
The experimental analysis has been performed by processing an interferometric dataset acquired over the Napoli bay,
a volcanic and densely urbanized area in Southern Italy that
includes the active caldera of Campi Flegrei, the Vesuvius volcano, and the city of Napoli. In particular, we consider the
overall time series of ENVISAT acquisitions collected by the
ASAR sensor from ascending orbits, which is composed by 64
SAR data. This dataset, which spans the 20032010 time interval and covers an area of about 100 100 km2 that corresponds
to an ENVISAT frame, is often used as a benchmark dataset
for DInSAR analyses [8], [14], [15], [18]. The selected dataset
has been processed by using the P-SBAS algorithm to generate the DInSAR products. In particular, in Fig. 3, the obtained
mean deformation velocity map is shown. This map has been
geocoded and afterward superimposed on a multilook SAR
image of the investigated area. The estimated mean deformation

velocity has been only computed in coherent areas; accordingly, areas in which the measurement accuracy is affected by
decorrelation noise have been excluded from the false-color
map. In particular, it is worth noting that in Fig. 3, a significant deformation pattern corresponding to the area of the
Campi Flegrei caldera is clearly shown. Moreover, the computation of the temporal evolution of the detected deformation
has also been carried out for each coherent point of the scene.
For instance, the chronological sequences of the computed
displacement of two specific points (the first located in the
maximum deforming area of the Campi Flegrei caldera while
the second in the Vomero residential hill which shows a slow
down lift movement) are plotted in the insets of Fig. 3. These
results are in accordance with ground truth measurements
[8], [10].
C. Experimental Results
As previously mentioned, a scalability analysis with respect
to the number of exploited computing nodes has been carried out both on the CNR-IREA cluster and on AWS Cloud.
Concerning the test performed at the CNR-IREA premises, the
speedup depicted as a function of the number of engaged nodes
is represented in Fig. 4. Such a plot shows the speedup ideal linear behavior (blue/diamonds), the Amdahls law (red/squares)
and the experimental results achieved on the above mentioned
cluster (green/triangles). The Amdahls law has been evaluated by computing the processing sequential fraction as the
ratio between the sum of elapsed times of the P-SBAS sequential parts and the total processing time on a single computing
node; it has turned out to be approximately 9% (fs = 0.09). It
is evident from Fig. 4 that the achieved speedup is definitely

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
8

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING

Fig. 4. P-SBAS performances on CNR-IREA cluster: Speedup as a function

of the number of nodes N (green/triangles). The ideal achievable speedup
(blue/diamonds) and the Amdahls law behavior (red/squares) are also shown.

Fig. 5. P-SBAS performances on AWS Cloud: Speedup as a function of

the number of nodes N (green/triangles). The ideal achievable speedup
(blue/diamonds) and the Amdahls law behavior (red/squares) are also shown.
Note that we refer to a node as to an AWS EC2 instance.

satisfactory; indeed the speedup curve is very close to the

Amdahls law with a slight deviation as approaching 16 nodes.
This result reveals that, by exploiting a HPC cluster, all the
factors that can hamper the efficiency, such as load unbalancing, communication times and I/O overhead [36], are basically
negligible at least up to 16 nodes. Hence the only significant
inefficiency source is the P-SBAS sequential fraction that is
taken into account by the Amdahls law and is still subject of
further improvements, being some sequential parts of the PSBAS processing chain currently on progress to be turned into
parallel ones.
Fig. 5, instead, shows the speedup that has been evaluated
through the scalable analysis carried out within AWS Cloud
by exploiting up to 16 instances. In this case, the processing
inherent sequential fraction has been estimated to be approximately 7.5% (fs = 0.075), thus being even slightly smaller
than in the cluster case. Such a difference is ascribable to
the fact that the processing elapsed times strongly depend on
the exploited specific computational environments that in the

cluster and Cloud cases are different; in particular, the processors of the Cloud nodes are slightly more powerful than those
of the cluster nodes (see Tables II and III). As Fig. 5 clearly
shows, also in this case, the speedup behavior is very close to
the Amdahls law and it begins to diverge as approaching 16
nodes. In Table IV, the Amdahls law and actual speedup values
evaluated in both the CNR-IREA cluster and AWS Cloud cases
are shown. Furthermore, the percentage deviations between the
theoretical (Amdahls law) and experimental behavior are provided, quantitatively confirming the good match between the
P-SBAS performance both on cluster and Cloud.
To provide an idea of the times as well as the economical expenses that P-SBAS took to complete the processing, in
Table V, the elapsed times relevant to the P-SBAS running
tests performed on both CNR-IREA cluster and AWS Cloud
by exploiting 1, 2, 4, 8, and 16 nodes are shown together with
the corresponding costs relative to the AWS usage. Note that
such costs consider both those relevant to the EC2 exploited
instances as well as those of the selected storage volumes.
On the contrary, the cost of the IDL licenses has not been
included because we used those available on the CNR-IREA
server with no additional expenses. By exploiting the AWS
Cloud, the P-SBAS processing times passed from 41 to less
than 7 h when moving from 1 to 16 nodes with a cost of USD
113 and 213, respectively. Table V shows that the P-SBAS parallel performances on Cloud are deemed satisfactory, as they
are absolutely comparable with those achieved on the dedicated
cluster.
The elapsed times and associated costs shown in Table V,
which are related to a maximum of 16 nodes, are undoubtedly
adequate when the processing of ENVISAT ASAR datasets on
the Cloud is concerned. However, in the perspective of dealing
with archives bigger than ENVISAT ones, as in the case of CSK
and Sentinel-1 data, the need of exploiting a larger number of
nodes is envisaged in order to keep the processing time of the
same order of magnitude. In this case, according to the performance behavior represented by the speedup curves of Fig. 5, the
discrepancy between the actual and the Amdahls law speedup
curves is expected to increase, thus significantly lowering the
efficiency.
In order to identify which is the performance bottleneck
when the number of parallel processes increases, some useful
metrics, such as the read/write bandwidth and average queue
length, provided by the AWS CloudWatch monitoring system
[20] have been analyzed in detail.
It turned out that the loss of efficiency relevant to the
16-nodes processing is ascribable to the I/O workload as it
understandably increases with the number of parallel processes
which concurrently read or write on the common storage volume. Hence, the factor that mainly lowers the P-SBAS scalable
performance in our case is essentially the storage volume
access bandwidth that is smaller than the network bandwidth
(256 MB/s vs. 10 Gb/s).
In the following the performances of two steps of the PSBAS algorithm, the image coregistration and the spatial phase
unwrapping (blocks C and I of Fig. 1), which are characterized
by very different I/O workloads, are thoroughly analyzed. To
this aim, we considered two metrics: the read/write bandwidth

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
ZINNO et al.: FIRST ASSESSMENT OF P-SBAS DINSAR ALGORITHM PERFORMANCES

TABLE IV
VALUES OF THE A MDAHL S L AW AND E XPERIMENTAL S PEEDUP ON B OTH CNR-IREA C LUSTER AND AWS C LOUD ;
T HEIR P ERCENTAGE D EVIATION IS A LSO S HOWN

TABLE V
P-SBAS P ROCESSING T IMES AND C OSTS

*Note that the reported costs include both the instances and the provisioned IOPS (SSD) volume
usage.

as well as the average length queue, both measured with respect

to the EBS storage volume. The former metric is provided by
AWS in KiB/s and has been converted in MB/s to be consistent
with the known storage volume access bandwidth, while the
latter metric represents the number of pending I/O requests
and, therefore, it is a latency measure. For the sake of simplicity in Tables VI and VII, only the metrics relevant to the
processing carried out with 4, 8, and 16 nodes are reported,
which are the most significant to understand how the I/O workloads affects the scalable performances when the number of
concurrent processes increases.
The image coregistration has been selected as it is one of
the most demanding step of P-SBAS in terms of I/O workload; moreover, it shows the poorest scalable performances
among all the P-SBAS algorithm steps. Indeed, accordingly to
Table VI, the image coregistration presents an efficiency that
strongly degrades while the number of nodes increases. Even
if this step could be liable to an improvement from a programming viewpoint in order to reduce its I/O workload, its
behavior is helpful to correlate the inadequate scalable performances to the read/write bandwidth saturation. As a matter of
fact, such a bandwidth is practically saturated for eight nodes
as it already reaches the amount of 230 MB/s of data transfer;
this value increases and becomes even 250 MB/s for 16 nodes.

The critical value of the average queue length for the employed
storage configuration is around 40 counts, indeed this number depends on the I/O capacity of the selected EBS volume
[20]. Once again eight nodes are already enough to reach the
maximum allowed latency threshold and a greater number of
employed nodes would not bring a significant reduction in the
elapsed time for this step.
On the contrary, the phase unwrapping step, even if very
burdensome from a computational viewpoint, is less heavy for
what concerns I/O operations. This step presents satisfactory
scalable performances as shown in Table VII, with a 70% efficiency in correspondence with 16 nodes and it would therefore
further scale if a greater number of nodes were used. Indeed, in
this case, both the read/write bandwidth and the queue length
values are evidently below the saturation threshold.
It is worth noting that the scalable performances of the PSBAS processing chain in the presented Cloud configuration
can be further improved by increasing the storage volume
access bandwidth by configuring a RAID 0 striping with a
greater number of provisioned IOPS volumes (up to 12).
This would allow us to exploit a larger number of nodes without saturating the storage volume access bandwidth, as long
as it is supported by an adequate network bandwidth, with the
performance limit given by the Amdahls law.

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
10

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING

TABLE VI
I MAGE C OREGISTRATION S TEP : E FFICIENCY AND I/O M ETRICS (R ETRIEVED F ROM THE C LOUDWATCH M ONITORING S YSTEM )
AS A F UNCTION OF THE N UMBER OF E XPLOITED N ODES ON AWS C LOUD

TABLE VII
P HASE U NWRAPPING S TEP : E FFICIENCY AND I/O M ETRICS (R ETRIEVED F ROM THE C LOUDWATCH M ONITORING S YSTEM )
AS A F UNCTION OF THE N UMBER OF E XPLOITED N ODES ON AWS C LOUD

V. C ONCLUSION AND F URTHER D EVELOPMENTS

In this paper, a first assessment of the scalable performances
of the P-SBAS DInSAR algorithm migrated to a CC environment has been presented. Such performances have been
evaluated, through the processing of a selected dataset acquired
by the ENVISAT ASAR system, by investigating the P-SBAS
algorithm experimental speedup on an increasing number of
nodes (up to 16) and by comparing it with the corresponding
theoretical speedup limit given by the Amdahls law. This analysis has shown a very good match between the Amdahls law
and the experimental speedup behavior that, in the worst case,
i.e., in correspondence with the 16-nodes test case, it presents
a discrepancy of about 17%. Moreover, the speedup obtained
within the AWS CC environment has been evaluated versus the
one achieved by exploiting an in-house HPC cluster, showing
comparable results and thus proving that the proposed Cloud
implementation does not introduce any significant overhead.
In summary, the P-SBAS scalable performances achieved
within a CC environment are definitely satisfactory for the considered number of employed nodes. Moreover, the evaluated
processing elapsed times show that it is possible to carry out
a P-SBAS elaboration of a typical ENVISAT SAR image time
series within 7 h with a cost of about USD 200. Hence, the
presented P-SBAS Cloud implementation is highly suitable to
process image time series acquired by the widely diffused longterm C-band SAR sensors generation, as for the ENVISAT
case. This result offers the possibility to simultaneously process many of such ENVISAT SAR datasets by exploiting a
sufficiently large number of nodes, which can be only available within CC environments, thus allowing us to process very
large portions of the ENVISAT archive in a short time.
However, in the perspective of the need to process massive SAR data amounts, as those already acquired by the

COSMO-SkyMed constellation as well as the ones expected

from the SENTINEL-1 satellites, the presented P-SBAS Cloud
implementation can present a drawback. In this case, indeed,
the use of a common NFS-based storage architecture (characterized by the concurrent reading and writing of all the parallel
processes on a single shared volume) is expected to become
a bottleneck. In fact, when a very large number of computing
nodes and, therefore, of parallel processes will be required by
the processing, both the disk access and the network bandwidth
would become saturated because of inevitable technological
limitations. In this regard, a solution based on a distributed file
system, properly designed to allocate data to be read or written
on local servers (thus minimizing the I/O bottlenecks and the
network workload) is a very promising answer to the need of
processing very large SAR data flows by scaling on a huge number of nodes. In this context, also the fault tolerance issues need
to be addressed in order to provide a P-SBAS implementation
which is robust with respect to nodes failures.
It is also worth noting that when a steady data stream
is expected, as in the Sentinel-1A case, the convenience to
add new scenes without reprocessing the whole stack (append
mode) should be assessed. Indeed, a deep analysis aimed at
evaluating the cost relevant to store, for an extended time frame,
the large additional amount of intermediate products, which are
needed to update the P-SBAS deformation time-series with the
new acquisitions, is required.
We further remark that, although the presented analysis is
fully focused on the AWS Cloud exploitation, the proposed
solution is also suitable to be deployed on other Cloud environments. Such a task is already under development; indeed,
the porting of the P-SBAS algorithm within the Helix Nebula
Cloud environment, which aims at provisioning a sustainable
European scientific Cloud [37], is a topic well worthy of further
investigation.

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
ZINNO et al.: FIRST ASSESSMENT OF P-SBAS DINSAR ALGORITHM PERFORMANCES

This work represents a relevant step toward the challenging EO scenario focused on the joint exploitation of advanced
DInSAR techniques and CC platforms for the massive processing of Big SAR Data. This will give the opportunity
of generating value added interferometric products on very
large scale and in short times, thus broadening the path to a
comprehensive understanding of Earths surface deformation
dynamics.

ACKNOWLEDGMENT
The authors would like to thank S. Guarino, F. Parisi, and
M.C. Rasulo for their technical support.

R EFERENCES
[1] E. Sansosti, F. Casu, M. Manzo, and R. Lanari, Space-borne radar interferometry techniques for the generation of deformation time series: An
advanced tool for Earths surface displacement analysis, Geophys. Res.
Lett., vol. 37, 2010, doi: 10.1029/2010GL044379.
[2] P. Berardino, G. Fornaro, R. Lanari, and E. Sansosti, A new algorithm
for surface deformation monitoring based on small baseline differential
SAR interferograms, IEEE Trans. Geosci. Remote Sens., vol. 40, no. 11,
pp. 23752383, Nov. 2002.
[3] F. Casu, M. Manzo, and R. Lanari, A quantitative assessment of the
SBAS algorithm performance for surface deformation retrieval from
DInSAR data, Remote Sens. Environ., vol. 102, pp. 195210, 2006.
[4] A. Manconi et al., On the effects of 3-D mechanical heterogeneities
at Campi Flegrei caldera, southern Italy, J. Geophys. Res. Solid Earth,
vol. 115, 2010, doi: 10.1029/2009JB007099.
[5] R. Lanari, F. Casu, M. Manzo, and P. Lundgren, Application of the
SBAS-DInSAR technique to fault creep: A case study of the Hayward
fault, California, Remote Sens. Environ., vol. 109, pp. 2028, 2007.
[6] F. Cal et al., Enhanced landslide investigations through advanced
DInSAR techniques: The Ivancich case study, Assisi, Italy, Remote Sens.
Environ., vol. 142, pp. 6982, 2014.
[7] R. Lanari, P. Lundgren, M. Manzo, and F. Casu, Satellite radar
interferometry time series analysis of surface deformation for Los
Angeles, California, Geophys. Res. Lett., vol. 31, 2004, doi:
10.1029/2004GL021294.
[8] M. Bonano, M. Manunta, A. Pepe, L. Paglia, and R. Lanari, From previous C-band to new X-band SAR systems: Assessment of the DInSAR
mapping improvement for deformation time-series retrieval in urban
areas, IEEE Trans. Geosci. Remote Sens., vol. 51, no. 4, pp. 19731984,
Apr. 2013.
[9] A. Pepe, E. Sansosti, P. Berardino, and R. Lanari, On the generation of ERS/ENVISAT DInSAR time-series via the SBAS technique,
IEEE Geosci. Remote Sens. Lett., vol. 2, no. 3, pp. 265269, Jul.
2005.
[10] M. Bonano, M. Manunta, M. Marsella, and R. Lanari, Long-term
ERS/ENVISAT deformation time-series generation at full spatial resolution via the extended SBAS technique, Int. J. Remote Sens., vol. 33,
pp. 47564783, 2012.
[11] S. Salvi et al., The Sentinel-1 mission for the improvement of the scientific understanding and the operational monitoring of the seismic cycle,
Remote Sens. Environ., vol. 120, pp. 164174, May 2012.
[12] A. Rucci, A. Ferretti, A. M. Guarnieri, and F. Rocca, Sentinel 1 SAR
interferometry applications: The outlook for sub millimeter measurements, Remote Sens. Environ., vol. 120, pp. 156163, May 2012.
[13] F. De Zan and A. Monti Guarnieri, TOPSAR: Terrain observation by
progressive scans, IEEE Trans. Geosci. Remote Sens., vol. 44, no. 9,
pp. 23522360, Sep. 2006.
[14] F. Casu et al., SBAS-DInSAR parallel processing for deformation timeseries computation, IEEE J. Sel. Topics Appl. Earth Observ. Remote
Sens., vol. 7, no. 8, pp. 32853296, Aug. 2014.
[15] P. Imperatore et al., Scalable performance analysis of the parallel SBASDInSAR algorithm, in Proc. IEEE Int. Geosci. Remote Sens. Symp.
(IGARSS14), Quebec City, Canada, pp. 350353, 2014.
[16] S. Elefante et al., G-POD implementation of the P-SBAS DInSAR algorithm to process big volumes of SAR data, in Proc. Conf. Big Data from
Space (BiDS14), 2014.

11

[17] P. Rosen, K. Shams, E. Gurrola, B. George, and D. Knight,

InSAR scientific computing environment on the cloud, in Proc.
AGU Conf., San Francisco, CA, USA, 2012 [Online]. Available:
http://fallmeeting.agu.org/2012/eposters/eposter/in31c-1508/
[18] S. Elefante et al., SBAS-DInSAR time series generation on cloud
computing platforms, in Proc. IEEE Int. Geosci. Remote Sens. Symp.
(IGARSS13), Melbourne, Australia, pp. 274277, 2013.
[19] M. Hansen et al., High-resolution global maps of 21st-century forest
cover change, Science, vol. 342, pp. 850853, 2013.
[20] Amazon. (2014). Amazon Web Services [Online]. Available: http://aws.
amazon.com/
[21] Google. (2014). Google Earth Engine [Online]. Available: http://www.
google.it/intl/it/earth/outreach/tools/earthengine.html
[22] HNX. (2015). Helix Nebula Marketplace [Online]. Available: http://hnx.
helix-nebula.eu/
[23] K. Yelick, S. Coghlan, B. Draney, and R. S. Canon, The Magellan report
on cloud computing for science, U.S. Dept. Energy, Office of Science,
Office of Advanced Scientific Computing Research (ASCR), Washington,
DC, USA, Dec. 2011.
[24] O. Terzo, L. Mossucca, A. Acquaviva, F. Abate, and R. Provenzano, A
cloud infrastructure for optimization of a massive parallel sequencing
workflow, presented at the 12th IEEE/ACM Int. Symp. Cluster Cloud
Grid Comput. (CCGrid12), Ottawa, Canada, 2012, pp. 678679.
[25] G. Franceschetti and R. Lanari, Synthetic Aperture Radar Processing.
Boca Raton, FL, USA: CRC Press, 1999.
[26] J. C. Curlander and R. McDonough, Synthetic Aperture Radar-System
and Signal Processing. Hoboken, NJ, USA: Wiley, 1991.
[27] E. Sansosti, P. Berardino, M. Manunta, F. Serafino, and G. Fornaro,
Geometrical SAR image registration, IEEE Trans. Geosci. Remote
Sens., vol. 44, no. 10, pp. 28612870, Oct. 2006.
[28] F. Gatelli et al., The wave-number shift in SAR interferometry, IEEE
Trans. Geosci. Remote Sens., vol. 32, no. 4, pp. 855865, Jul. 1994.
[29] A. Pepe and R. Lanari, On the extension of the minimum cost flow algorithm for phase unwrapping of multitemporal differential SAR interferograms, IEEE Trans. Geosci. Remote Sens., vol. 44, no. 9, pp. 23742383,
Sep. 2006.
[30] V. Andrikopoulos, T. Binz, F. Leymann, and S. Strauch, How to adapt
applications for the cloud environment, Computing, vol. 95, pp. 493
535, 2013.
[31] R. Sandberg, D. Goldberg, S. Kleiman, D. Walsh, and B. Lyon, Design
and implementation of the SUN network filesystem, in Proc. USENIX
Conf., pp. 119130, 1985.
[32] Exelis. (2014). Exelis Visual Information SolutionExelis Interactive
Data Language (IDL) [Online]. Available: http://www.exelisvis.com
[33] Globesar AS. (2015). GlobesarDiscovering Global Changes [Online].
Available: http://www.globesar.com/
[34] H. El-Rewini and M. Abd-El-Barr, Advanced Computer Architecture and
Parallel Processing. Hoboken, NJ, USA: Wiley, 2005.
[35] G. Hager and G. Wellein, Introduction to High Performance Computing
for Scientists and Engineers. 2010.
[36] I. Foster, Designing and Building Parallel Programs. USA: Addison
Wesley, 1995.
[37] Helix Nebula. (2014). Helix NebulaThe Science Cloud [Online].
Available: http://www.helix-nebula.eu

Ivana Zinno was born in Naples, Italy, on July 13,

1980. She received the Laurea (summa cum laude)
degree in telecommunication engineering and the
Ph.D. degree in electronic and telecommunication
engineering from the University of Naples Federico
II, Naples, Italy, in 2008 and 2011, respectively.
In 2011, she received a grant from the University
of Naples to be spent at the Department of Electronic
and Telecommunication Engineering for research in
the field of remote sensing. Since January 2012, she
has been with IREA-CNR, Napoli, Italy, where she
is currently a Post-doc Researcher, mainly working on the development of
advanced differential SAR interferometry (DInSAR) techniques for deformation time series generation by also exploiting parallel computing platforms.
Her research interests include microwave remote sensing, differential SAR
interferometry applications for the monitoring of surface displacements and
information retrieval from SAR data by exploiting fractal models and techniques.

Thisarticlehasbeenacceptedforinclusioninafutureissueofthisjournal.Contentisfinalaspresented,withtheexceptionofpagination.
12

IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING

Stefano Elefante received the Laurea degree (summa

cum laude) from the University of Naples, Napoli,
Italy and the Ph.D. degree in aerospace engineering
from the University of Glasgow, Scotland, U.K., in
1997 and 2001, respectively. He holds a PgCert in
financial engineering from the Columbia University,
New York, NY, USA, and in applied statistics from
the Birkbeck College, University of London, London,
U.K., since 2008 and 2011, respectively.
He has more than 10 years of diverse professional
experience in academia and industry, during which
he has been involved in different scientific fields from statistical genetics and
financial mathematics to aerospace technologies. From 2005 to 2009, he was a
System Analyst and Research Engineer with Boeing Research & Technology
Europe, Madrid, Spain, where he conducted extensive research in the field of
air traffic management developing and implementing aerospace systems for
improving airport and airspace efficiency. Since 2011, he has been holding a
research position with IREA-CNR, Napoli, Italy. His research interests include
investigating innovative mathematical methodologies for remote sensing applications. He is currently developing novel parallel algorithms for synthetic
aperture radar interferometry (InSAR) within cluster, grid and cloud computing
environments.

Olivier Terzo was born in Dijon, France, on

September 1969. He received the Degree in electrical engineering technology and industrial informatics from the University Institute of Nancy, Nancy,
France, the M.Sc. degree in computer engineering and the Ph.D. degree in electronic engineering
and communications from the Polytechnic of Turin,
Turin, Italy.
He is a Senior Researcher with the Istituto
Superiore Mario Boella (ISMB), Torino, Italy, from
2004 to 2009. He worked in the e-security laboratory, mainly with a focus on P2P protocols, encryption on embedded devices,
security of routing protocols and activities on grid computing infrastructures,
and from 2010 to 2013, he was the Head of the Research Unit Infrastructure
Systems for Advanced Computing (IS4AC) at ISMB. Since 2013, he is the
Head of Research Area: Advanced Computing and Electromagnetics (ACE),
dedicated to the study and implementation of computing infrastructure based
on virtual grid and Cloud computing and to the realization of theoretical and
experimental activities of antennas, electromagnetic compatibility, and applied
electromagnetics. He has authored about 60 papers in conferences, journals
and book chapters and he is the editor of the book Cloud Computing With eScience Applications (CRC Press, 2015). His research interests include hybrid
private and public cloud distributed infrastructure, grid and virtual grid, mainly
activities are on applications integration in cloud environments.

Lorenzo Mossucca was born in Torino, Italy, on

February 1982. He received the Laurea degree in
computer engineering from the Polytechnic of Turin,
Turin, Italy, in 2014.
From 2007, he works as a Researcher with
the Istituto Superiore Mario Boella (ISMB) in
the Infrastructures and Systems for Advanced
Computing (IS4AC) research unit. He has authored
and coauthored about 30 international journal, book
chapter, and conference papers. He joins in Technical
Program Committee and as a Reviewer for many
international conferences and editor of the book Cloud Computing With eScience Applications (CRC Press, 2015). His research interests include studies
on distributed databases, distributed infrastructures, grid and Cloud Computing,
migration of scientific applications to cloud, bioinformatics, and earth sciences
fields.

Riccardo Lanari (M91SM01F13) graduated in

electronic engineering (summa cum laude) from the
University of Napoli, Federico II, Napoli, in 1989.
In the same year, following a short experience
with ITALTEL SISTEMI SPA, he joined IRECE
and after Istituto per il Rilevamento Elettromagnetico
dellAmbiente (IREA), a Research Institute of the
Italian National Research Council (CNR), Napoli,
where, since November 2011, he is the Institute
Director. He has lectured in several national and
foreign universities and research centers. He was
an Adjunct Professor of electrical communication with the lUniversit del
Sannio, Benevento, Italy, from 2000 to 2003, and from 2000 to 2008, he
was the Lecturer of the Synthetic Aperture Radar (SAR) module course of
the International Master in Airbone Photogrammetry and Remote Sensing
offered by the Institute of Geomatics, Barcelona, Spain. He was a Visiting
Scientist at different foreign research institutes, including the Institute of
Space and Astronautical Science, Japan, in 1993; German Aerospace Research
Establishment (DLR), Germany, in 1991 and 1994; and Jet Propulsion
Laboratory, Pasadena, CA, in 1997, 2004, and 2008. He is the holder of two
patents, and he has authored or coauthored 80 international journal papers and
the book Synthetic Aperture Radar Processing (CRC Press, 1999). His research
interests include SAR data processing field as well as in SAR interferometry
techniques.
Mr. Lanari is a Distinguished Speaker of the Geoscience and Remote Sensing
Society of IEEE, and he has served as the Chairman and as a Technical Program
Committee Member at several international conferences. Moreover, he acts as a
Reviewer of several peer-reviewed international journals. He received a NASA
Recognition and a Group Award for the technical developments related to the
Shuttle Radar Topography Mission.

Claudio De Luca was born in Naples, Italy, on July

16, 1987. He received the Laurea degree (110/110) in
telecommunication engineering from the University
of Naples Federico II, Naples, Italy, in 2012. He is
currently pursuing the Ph.D. degree in computer and
automatic engineering at the same university.
His research interests include cloud computing
solution for intensive processing of remote sensing data, development of advanced algorithm for
Sentinel-1 SAR and InSAR data processing.

Michele Manunta was born in Cagliari, Italy, in

1975. He received the Laurea degree in electronic
engineering and the Ph.D. degree in informatics
and electronic engineering from the University of
Cagliari, Cagliari, Italy, in 2001 and 2009, respectively.
From 2002, he has been with the Istituto per
il Rilevamento Elettromagnetico dellAmbiente
(IREA), an Institute of the Italian National Research
Council (CNR), where he currently holds a
Researcher position. He was a Visiting Scientist
at the Institut Cartografic de Catalunya, Barcelona, Spain, in 2004, and the
Rosenstiel School of Marine and Atmospheric Science of the University of
Miami in 2006. He has been collaborating in various national and international
initiatives for the exploitation of satellite technologies, and in particular of SAR
techniques. His research interests include high resolution SAR and DInSAR
data processing and application, Cloud and GRID computing exploitation for
SAR interferometry applications.

Francesco Casu received the Laurea (summa cum

laude) and Ph.D. degrees in electronic engineering
from the University of Cagliari, Cagliari, Italy, in
2003 and 2009, respectively.
Since 2003, he has been with the IREA-CNR,
Napoli, Italy, where he currently holds a permanent
Researcher position. He was a Visiting Scientist at the
University of Texas at Austin, Austin, TX, USA, in
2004, the Jet Propulsion Laboratory, Pasadena, CA,
USA, in 2005, and the Department of Geophysics,
Stanford University, Stanford, CA, USA, 2009.
Moreover, he acts as a Reviewer of several peer-reviewed international journals. His research interests include DInSAR field, multipass interferometry
(particularly concerning the improvement of the SBAS-DInSAR algorithm),
SBAS-DInSAR measurement assessment, with particular emphasis on novel
generation satellite constellations such as COSMO-SkyMed, TerraSAR-X and
Sentinel-1, development of DInSAR algorithms for unsupervised processing of
huge SAR data archives by exploiting high-performance computing platforms,
such as GRID and Cloud computing ones.